Augmented reality allows interaction between users and real-world objects and virtual or computer-generated objects and information. These interactions may include verbal, physical, or other input signals made within a scene such as a room. However, in some scenes, monitoring and processing the data from the entire volume of the scene may be impractical where scan resolution is high, the physical environment is large, limited resources are available, and so forth. What is desired is a way to focus scanning and processing resources on particular regions of interest within the scene to reduce resource requirements.